Thinking N = 1
Understanding and applying individual variability in research results
In a recent post, I addressed the importance of defining and distinguishing statistical significance – whether or not a study outcome is likely due to chance – and clinical significance – whether or not a study outcome is impactful when applied to real life scenarios. In addition to gauging the legitimacy and meaningfulness of study data on a population level, it’s important to consider how an intervention impacts our own outcomes, which is commonly referred to as an N =1 perspective — with, “N” representing the sample size of a given study. After all, the most important question when it comes to applying research results to our personal choices is, “How does this play out for ME in real life?” — or, “how does this look when N = 1?”
In this regard, especially in the world of health and fitness where responses to lifestyle interventions often vary widely, individual variability is a critical factor. And, I’ll bet that subject results are probably more varied than you expect. Here, I’ll share an example of one such instance, and discuss my current framework for leveraging population data while remaining adaptive in my own health and fitness habits.
As I just alluded to, it’s important to acknowledge that most of the time we are discussing study results, we are referencing some form of an average of the study population’s outcomes, whether it be an effect size, risk ratio, or mean difference. This is useful for determining how a given intervention – say an exercise protocol or a diet modification – tends to impact people directionally, with some insights on the magnitude of the impacts; however, due to human variation across subjects, these population outcomes may not perfectly line up with our individual responses to an intervention. In fact, population statistics may directly oppose the directionality and magnitude of individual outcomes in some cases.
One prime example is Schoenfeld et al.’s study on how training volume impacts strength and hypertrophy outcomes, where subjects conducting 1, 3, or 5 sets per exercise per session showed a dose-response relationship with more sets leading to more muscle growth according to the groups’ mean changes in muscle size. Interestingly, though the average changes in muscle thickness display a clear step-wise pattern, looking at individual subjects’ responses reveals some subjects in the 1 set per exercise group gained more muscle than some subjects in the 3 and 5 sets per exercise groups. In addition, for some muscle groups, some subjects in all three groups showed zero change in muscle thickness or even a slight decrease in size. This is visually apparent in Figure 3 of the study, and even more obvious in the scatter plots in this article where James Krieger, one of the study’s authors, references the outcomes.
So, does this mean that the study’s conclusion is bologna somehow driven by a gym industry conspiracy to make greater profits by keeping us in the weight room longer? Well, no. Greater volume did lead to greater hypertrophy on average for the study population, it just didn’t lead to greater hypertrophy for every individual subject across the board. And, the degree to which volume was related to muscle growth varied from one subject to another. If you go back and look at the numbers, you may be shocked by the magnitude of this variation, but I found many similar data distributions while researching for the Muscle Mania Series.
To further wrap your head around this idea of individual variation, take a look at Figures 1 and 2 below. The first plot portrays what one might perceive all study results to look like based on the way we commonly discuss and apply effect sizes, risk ratios, and means. The second graph depicts the reality of a lot of study results, in which individual outcomes vary, with some subjects outperforming the mean, some underperforming, and others falling in between.
With this in mind, rather than following research results strictly and adopting study protocols to a, “T”, I’ve shifted towards using research data as guidelines with which I can test and monitor my personal responses to different diets, training protocols, etc.. Importantly, this involves using objective markers – like blood pressure, heart rate, and body weight, for example – to guide my choices according to what research data suggests about these endpoints and risk factors.
For example, elevated saturated fat consumption is famously correlated with increased cardiovascular disease risk; however, this is only true past a certain threshold of consumption, with elevated Apo-B count being the causal mechanism that ultimately leads to the cardiovascular disease. Since the threshold at which saturated fat consumption raises Apo-B levels varies widely from one person to the next, the most logical approach is to track one’s Apo-B levels, adjust saturated fat consumption accordingly if necessary, and continue to monitor and adapt over time. In this way, we are able to tailor our decisions to modify our individual risks and outcomes.
I first fully grasped this perspective – following research results as general guidelines and adapting according to individual responses – after listening to Dr. Brad Schoenfeld, the first author from the example study above, espouse this approach in relation to his work and life. Interestingly, this means that steering in the opposite direction of study results can actually be the most evidence-based approach if you observe your individual responses deviating from those in research.
Ultimately, the most important outcome at the individual level is my own. For this reason, I place the utmost value in my own data – whether it be my weight, heart rate, run time, strength, or blood markers – while utilizing research data as a guide to potentially valuable interventions for me to try. In addition, I leverage data around different metrics (ex: Apo-B correlating to cardiovascular disease) to gauge whether I’m moving in the desired direction. Because, from this N= 1 perspective, we all really are special in our own ways.